Abstract
N6-methyladenosine (m6A) is a pervasive mRNA modification that regulates RNA fate through effects on RNA-protein interactions, stability and translation. We previously showed that replication of human betacoronaviruses, OC43 (hCoV-OC43) and SARS-CoV-2, is sensitive to depletion or pharmacological inhibition of the m6A RNA methyltransferase METTL3, resulting in reduced viral RNA and protein accumulation. In other viral systems, such antiviral effects have been attributed to enhanced interferon (IFN) signalling and interferon-stimulated gene (ISG) induction. Here, using hCoV-OC43 we show a requirement for METTL3 that is independent of canonical IFN responses. Pharmacological inhibition of METTL3 with STM2457 failed to potentiate type I IFN signalling, global ISG expression, or the non-canonical inflammatory transcriptional programme associated with OC43 infection. Furthermore, pathogen-associated RNA sensing by RIG-I or MDA5 is not required for the antiviral effect of the STM2457. ISGs reported to be most potently antiviral against OC43 are either not significantly induced by METTL3 inhibition during infection or are not required for the antiviral activity. Nevertheless, defects in viral gene expression and progression through the viral life cycle are detectable within 6 h of STM2457 treatment and host cell transcription is dispensable for STM2457 antiviral activity. Lastly, a METTL3-directed Proteolysis Targeting Chimera (PROTAC) phenocopied STM2457, producing IFN-independent antiviral activity and ruling out off-target inhibition of viral RNA methyltransferases as a plausible explanation. Together, these findings define a direct, proviral role for METTL3 in coronavirus infection consistent with a model in which METTL3-catalysed m6A modification of viral RNA is required for efficient viral life cycle progression.